The Role of X Inactivation and Cellular Mosaicism in Women's Health and Sex-Specific Diseases

A perusal of publications devoted to women's health issues suggests that women's diseases are those diseases affecting the breast and ovaries, menstrual disorders, pregnancy, infertility, menopause, postmenopausal osteoporosis, and other disorders attributable to sex differences in the reproductive apparatus. Recently, the list has broadened to recognize that men and women differ in response to their diseases. Such sex-specific responses may impede the ability of physicians to diagnose and appropriately treat a host of diseases, including cardiovascular disorders, infectious diseases, and psychiatric disorders. Also included are the differences in response to pharmaceuticals based on the documented sex differences in the kinetics of drug metabolism. Nonetheless, this expanded view of what constitutes a sexual dimorphism in the expression of disease often fails to consider biological reasons for the sex-specific manifestations other than hormones or lack of them.

Certainly, gonadal and hormonal differences, testosterone in men and estrogen in women, are responsible for some sex differences in disease. However, although hormones should not be ignored, neither should they dominate the view of sex differences. Even when environmental and hormonal differences between the sexes are not prominent, it is clear that mortality and morbidity are greater in males. Table 1 shows that more males than females die in infancy and preschool periods. The infant mortality is greater in males (averaging about 20% higher), irrespective of the length of gestation. Based on studies of recognized fetal loss, the greater loss of males is also observed in utero.

What then is responsible for the greater biological vulnerability of males? At least some of the sex difference in vulnerability is due to males having a single X chromosome and females having 2 X chromosomes; compensatory mechanisms having arisen to equalize the sex difference in the number of X chromosomes; and females being mosaics.

The sex chromosomes in females consist of a pair of X chromosomes, whereas males have a single X chromosome partnered with a Y chromosome. The Y chromosome carries the critical determinants of maleness. However, it carries little else, having lost most of its genetic content during mammalian evolution. Whereas more than a thousand genes reside on the X chromosome (X-linked genes), the Y chromosome carries very few functional genes (probably <100) and lacks working copies of most of the X-linked genes.¹ It is likely that many of the sex differences in disease manifestations are attributable to the sex differences in the number of X chromosomes. Having only 1 copy of X-linked genes (1 allele) makes males more vulnerable to deleterious mutations that adversely affect the function encoded by these genes, certainly more vulnerable than females with 2 copies (2 alleles). If his mutated allele is defective, a male cannot perform the function encoded by that gene. Yet the same mutated allele is usually less deleterious to a female, because she has a normal functioning copy (on the other X chromosome). This is why so many male-only diseases are attributable to defective genes on the X chromosome.

Among the proteins encoded by the X chromosome are antibodies that protect against infections. A deficiency of these proteins causes many X-linked immunological diseases (for example, Bruton agammaglobulinemia and Wiskott-Aldrich syndrome) (Table 2).² Clearly, mutations that compromise the function of such proteins contribute to the sex differences in the incidence of septicemia and meningitis of the newborn.^{3 - 4} Males, with only 1 allele, are compromised more than females, who are protected by their second copy of the gene. Although mutations leading to a severe antibody deficiency may not permit survival past childhood, less lethal reductions in the immunoglobulin levels could increase susceptibility to bacterial infections in adults.

Immunodeficiencies are not the only cause of male vulnerability. Single mutations in many of the 1100 X-linked genes can generate a deficiency of a host of essential proteins in males. For many of the eclectic group of disorders caused by X-linked mutations (some shown in Table 2), males usually have a more severe form of the disease than females. Males in contrast with females manifest full-blown Duchenne muscular dystrophy, hemophilia, and Lesch-Nyhan syndrome, and they die in utero or in early infancy of hyperammonemia (ornithine transcarbamylase deficiency) and incontinentia pigmenti. In contrast, females with the same mutant alleles may have no recognizable clinical abnormalities (like most females who carry Lesch-Nyhan syndrome or Duchenne mutations). Alternatively, females may have milder symptoms than their male relatives (like female carriers of ornithine transcarbamylase deficiency and incontinentia pigmenti). This female advantage is due to their having a normal copy of the gene, along with the mutated copy.

The opportunity to carry 2 alleles at the same locus provides an advantage for females, even though both alleles do not function in the same cell. In fact, like males, females effectively have a single X chromosome as they have only 1 working copy of a gene in each of their cells. Because of X inactivation—the mechanism that equalizes the sex differences in numbers of X chromosomes—the number of working X chromosomes in normal females is reduced from 2 to 1 during embryogenesis.^{5 - 6} Yet, having only a single functional X chromosome in each cell has different implications for the health of males and females, because normally only females are mosaics.

Females are mosaics because X inactivation creates 2 populations of cells differing in the parental origin of the active X (discussed in Migeon⁶ ). As the active X is usually chosen randomly, either chromosome has a chance to be the working X. Therefore, a female has a mixture of cells in all her tissues, some cells expressing the alleles on the X inherited from her mother, the other cells expressing the paternal alleles. The ratio of one cell population to the other may differ among individuals, and even among tissues of the same individual, due to stochastic events or rare mutations affecting the choice process.^{7 - 8} However, most females are in fact mosaics, at least during early embryogenesis when X inactivation occurs.

For the human female, cellular mosaicism is not an abstract concept. Because many X-linked genes have multiple alleles, the alleles on a female's 2 X chromosomes are likely to differ. When a deleterious mutation in 1 allele disables the cells expressing it, females have normal cells to perform the compromised function. Although the distribution of the 2 cell populations and intermingling of these cells may vary, the mosaic patches are small enough so that a female usually has a greater variety of gene products in all her tissues than a male does. The bottom line is that although females have only a single working set of X-linked genes in each cell, they have a backup copy in reserve.

Although the 2 clonal cell populations appear discrete, cells interact with each other in myriad ways. Such interactions usually foster a kind of metabolic cooperation between them.^{9 - 10} Because of metabolic cooperation, women with 1 copy of the mutant allele that is harmful or even lethal for their sons often show no deleterious effects. Their normal cells may provide enough of the essential gene product to correct the defect in the mutant cells, or at least enough to circumvent the lethality of the mutation. Cell-to-cell transfer of gene products masks the genotype, because the mutant cells are not really deficient. This occurs in the X-linked lysosomal diseases like Hunter and Fabry syndromes caused by the deficiency of lysosomal enzymes involved in the metabolism of large intracellular proteins^{11 - 12} (Table 2). Such enzymes freely enter and exit the lysosomes, the site of their digestive activity, and are transferred from one cell to another by mannose-6-phosphate–mediated endocytosis. When the enzyme is deficient, the undigested products accumulate and plug up or otherwise disrupt the normal function of the lysosomes. Iduronate sulfatase, the enzyme defective in Hunter syndrome, is synthesized in some cells of carrier females but not in others. The cells that make and secrete this enzyme transfer it to the cells that cannot; in this way, the defect in the mutant cells is corrected. This transfer also occurs in carriers of Fabry disease, but because α-galactosidase A is a lower-uptake enzyme, it is transferred less efficiently than iduronate sulfatase. As a consequence, Fabry heterozygotes may manifest some of the deleterious effects of this mutation; even then, their disease is usually much milder than that of affected males.¹² Metabolic cooperation also occurs in many tissues of Lesch-Nyhan heterozygotes, who benefit from the transfer of small nucleotides from normal to mutant cells via gap junctions^{9 - 10} (Table 2).

Not all interactions are favorable to females. Sometimes the products in cells expressing the mutant allele interfere with the function of the cells expressing the normal allele. This kind of interaction occurs in the craniofrontonasal syndrome, a developmental disorder that leads to premature fusion of the coronal suture of the skull.¹³ During embryonic development, cells in one lineage may induce function in another lineage or may actually repel the physical movement of the other. This happens in the development of the skull, where the growth of bones must be constrained in some way so that they do not prematurely fuse with one another, constricting the growth of the brain. Craniofrontonasal syndrome is caused by a deficiency of ephrin B1 (EPNB1), an X chromosome–encoded member of the ephrin family of transmembrane proteins, which are very important signaling molecules in many developmental processes.¹³ The role of EPNB1 is to define the position of the coronal suture of the skull, and it is thought that this is done by inhibiting cells from crossing the normally sharp neural crest-mesoderm tissue boundary.

Paradoxically, EPNB1 mutations produce more severe defects in heterozygous females than in hemizygous males.¹³ Males have few, if any, abnormalities in contrast with the carrier females who have full-blown craniosynostosis and other cranial abnormalities (Table 2). Why do mutations in EFNB1 produce cranial abnormalities only in females? The prevailing thought is that the mixture of mutant and normal cells in some way perturbs the signaling process required for the formation of the future coronal suture and, as a consequence of faulty signaling, the bones prematurely fuse. To explain why males with the mutation are so minimally affected, it has been suggested that other functionally redundant members of the ephrin family substitute for EFNB1 in completely deficient cells, but cannot do so in the female because of the admixture of mutant and wild-type cells. Although relatively rare, cellular interference of this kind may cause problems in females that are not observed in hemizygous males. Craniofrontonasal syndrome might be considered a female-only disorder because of the sex differences in pathogenesis and clinical manifestations.

Some X-linked mutations adversely affect the growth of cells, impeding their ability to utilize nutrients, or respond to growth signals. The presence of 2 kinds of cells, differing in their proliferative capacities, sets up a competition between them.⁷ Therefore, another interaction between the cell populations in the mosaic female is growth competition. The cells that reproduce faster will eventually outgrow the others; even small differences in growth rate have an effect. Cells expressing mutant alleles that interfere with proliferation are likely to be eliminated in time by overgrowth of the cells expressing the normal allele. The rate of elimination can be rapid (eg, the death of mutant cells in incontinentia pigmenti or the failure of mutant T cells to migrate to the marrow in Wiskott-Aldrich syndrome) or mutant cells can be lost gradually (eg, the enzyme-deficient blood cells in Lesch-Nyhan syndrome).⁷

The elimination of mutant cells leads to what is called unbalanced or skewed X inactivation. Approximately 10% of females have skewing such that greater than 95% of their cells express the same parental allele. The cells that are lost are often mutant, but even normal cells may be lost for stochastic reasons. This skewing may have no clinical significance for some of these females, as there are no clinically relevant mutations being expressed, or they may be masked on the inactive X. However, in some cases the skewing, no matter its cause, will influence the expression of a genetic disease.

Skewing that favors the normal allele may be caused by death or failure of mutant cells to migrate to appropriate tissues and is associated with no clinical abnormalities. When mutant cells have no proliferative disadvantage, clinical manifestations may occur if having only 50% normal cells is not sufficient, as in ornithine transcarbamylase deficiency or Rett syndromes (Table 2). In any case, the severity of clinical symptoms in any X-linked heterozygote will be influenced by skewing, no matter which allele is favored or the cause of the skewing. Skewing that favors the mutant gene will increase the manifestations, whereas skewing favoring the normal allele will decrease the likelihood of symptoms.

In some females, 1 mutant allele is enough to produce clinical symptoms; however, their disease will be less severe than in males. For example, if the mutation is lethal to males in utero, females may have some clinical manifestations. However, they were protected against death in utero by their normal copy of the gene (Table 2). Clearly, these sex-specific diseases are an outgrowth of the sex difference in numbers of X chromosomes. Incontinentia pigmenti and Rett syndrome are female-only diseases because the disease in males is usually lethal in utero or has different manifestations. Such disorders are often thought to affect only 1 sex, because the sex differences in expression of X-linked mutations is not always recognized. The practice of defining disorders by their syndromic phenotypes has proven to be enormously misleading as to the cause of the disease. The otopalatodigital syndrome group of bone malformations was considered as 4 independent X-linked diseases until the filamin A gene was shown to be mutant in all of them¹⁴ (Table 2). The manifestations differ according to the site in the gene that is mutated and the sex of the individual in whom it occurs. Rett syndrome was thought to affect only females until identification of the methyl CpG protein 2 gene (MeCP2) showed that the same mutation had different manifestations in males. In fact, for most X-linked disorders, females and males have their own form of the disease, based on the presence or absence of cellular mosaicism.

Mosaicism itself contributes to disease heterogeneity among females. If and how they manifest a disorder depends on the makeup of the mosaic cell populations and the nature of all the other alleles on their active X chromosomes. The composition of the mosaicism is determined initially by the random choice of active X during early embryogenesis.⁶ Only after X inactivation creates the cellular mosaicism does the influence of all the alleles on the 2 X chromosomes come into play. The ultimate composition of the mosaic cell populations is determined by the dynamic interaction between genes on the 2 X chromosomes. The 2 populations of cells, each one commandeered by a different active X chromosome, take part in a virtual competition for predominance. In most females, the result of the competition ends in a dead heat, one gene canceling out any minor selective advantage of another and both populations are fairly equally represented in every one of their tissues.

In a number of females, the result of the competition produces a winner and one cell population becomes preeminent in the tissues in which the more influential genes are expressed. In most cases, the cellular selection mechanism weeds out the weaker cells. Whether or not cell selection occurs is determined by multiple factors. Some factors have to do with the nature of the mutation, such as the amount of protein remaining in the mutant cells. Other factors include the ability of the gene product to be transferred from wild-type to mutant cells, and the degree to which it affects cell proliferation. Skewing is most apparent when the product of the mutant gene is nontransferable and has a strong influence on cell proliferation. In addition, other genes can affect the result; even the strong influence of mutant alleles at one X-linked locus can be modified by those at another. Therefore, the outcome of the competition between the 2 cell populations may differ from one tissue to another, depending on expression patterns of relevant genes.

Some females lose the biological advantage. Occasionally, the mutation confers a proliferative advantage rather than disadvantage and leads to cell selection favoring the mutant allele. Because skewing favors their mutant cells, many heterozygotes for adrenoleukodystrophy (Table 2) manifest the disease known as adrenomyeloneuropathy, a milder disorder that affects the spinal cord rather than the brain.^{15 - 16} If the competition between cell populations favors the mutant cells, females may manifest diseases that are usually observed only in males.¹⁷ Such females are termed manifesting heterozygotes because they express the male version of the disease despite having only 1 copy of the mutation. Other causes of manifesting heterozygotes include the nature of the chromosome on which the mutation resides. In some females, chromosome abnormalities involving the X chromosome eliminate the cellular mosaicism and unmask mutant alleles on the X chromosome, which is consistently active, or mask normal alleles that are never expressed.¹⁸ In rare cases, females are not mosaic because of mutations affecting the initial process that chooses the active X chromosome; one parental chromosome had an edge over the other.^{19 - 20} Furthermore, some monozygotic twins are prone to extreme skewing.^{21 - 22} Clearly, there are many determinants of a female's X-linked phenotype. Although mosaicism affords a significant biological advantage, the outcome is never certain. It depends on many dynamic interactions that are influenced by chance events as well as the variations in the normal blueprint for development.

Up to now the effect of the single active X and cellular mosaicism on the manifestations of X-linked diseases has been considered. However, the influence of X inactivation is not limited to the X chromosome, as any disease process with an X-linked component might be influenced by patterns of X inactivation and this possibility should be considered for any disorder that occurs more frequently in one sex than the other, or has different manifestations in one sex than the other.

One class of diseases that is much more prevalent in females than males is autoimmune disease. These include arthritis and thyroiditis in which antibodies to the relevant tissues are found in the blood of the affected individuals. Whereas, type 1 diabetes mellitus is the only major organ-specific autoimmune disorder not to show a strong female bias, diseases like thryroiditis, systemic lupus erythematosus, and scleroderma show 3- to 10-fold more affected females. For example, females comprise 90% of those individuals who develop systemic lupus erythematosus. Although the female prevalence is often attributed to the effect of estrogen, Stewart²³ pointed out that other sex differences might have as much or more relevance to autoimmune disease, for example, X inactivation.

A unique feature of autoimmune disorders is the apparent loss of immunological tolerance to self-antigens. In females, the self-proteins expressed and presented for tolerization differ in the 2 cell populations. Because this mosaicism is also present in the tolerizing cells of the thymus, the sampling process involved in recognizing self might miss the antigens on one of the parental X chromosomes. The probability that a self-antigen encoded by the X chromosome might elude the recognition process would be increased if X inactivation were highly skewed, such that one of the parental X chromosomes was expressed in relatively few cells. The more extreme the skew in thymic cells (specifically, the dendritic cells that present antigen), the more likely it is that antigens will evade the tolerizing process.

Recent studies lend support to this conjecture. Scleroderma is an autoimmune disease of connective tissue whose etiology is unknown and not thought to be X-linked, or even to have a strong genetic component. Therefore, it is surprising that nearly half of a group of 55 females with scleroderma have extremely skewed X inactivation in their blood cells.²⁴ It is likely that scleroderma is the consequence and not the cause of the skewing. If skewing occurs for stochastic or other reasons, the chance is increased for antigens expressed on the minor population of cells to escape recognition. One would not expect all affected individuals to show skewing, as it is likely that there are other risk factors for this disease. Also, evidence is accumulating that a third or more of individuals with the kind of thyroid disease associated with thyroid antibodies show skewed patterns of inactivation.^{25 - 26} On the other hand, in the scleroderma study, skewing in blood cells was not observed in individuals with rheumatoid arthritis, so this may not be a factor in all autoimmune diseases.

Most of our knowledge of human biology has been acquired by viewing it through the window of disease or deleterious mutations. However, it is likely that the contribution of cellular mosaicism to sex differences is not limited to disease. Studies of educational performance show that from the first days of school, girls outperform boys, are more attentive, and are more persistent at tasks.²⁷ Recent studies have shown that sex affects the way a person's brain responds to humor.²⁸ It does not seem far-fetched to think that cellular mosaicism may have a role in some of these sex differences in behavior. Even in absence of disease alleles, a female is a composite of 2 intermingling cell populations that are sharing gene products with one another. Clearly, mosaicism based on X inactivation has the potential to generate cellular diversity for many physiological processes. Based on observations of color vision, Smallwood et al²⁹ pointed out that cellular diversity is advantageous as a general strategy for enhancing the efficiency of signal processing and transmission. For example, in new world monkeys, the single X-linked pigment gene has 3 alleles, each encoding a pigment with a different spectral sensitivity. Males and homozygous females have dichromatic color vision, whereas heterozygous females have trichromatic color vision, and this is associated with enhanced chromatic discrimination.²⁹ Just as the ability to express a variety of normal color vision alleles enhances the way that color is perceived, having cells that collaborate on the elaboration of a protein may result in novel molecules, and such molecules may enhance the function being performed. The diversity provided by expressing 2 different alleles simultaneously, yet in different cells, is certain to lead to novel effects.

In conclusion, many sex-specific manifestations of disease are the consequence of both sexes having a single working X chromosome. Mutations of an X-linked gene in males leads to male-only diseases, because they have only the defective copy of the gene, whereas females have a normal copy in reserve. Mosaicism for the X chromosome proteome plays a large role in determining a female's disease response. Most often, mosaicism is advantageous, mediating transfer of essential gene products from normal to mutant cells, or elimination of deleterious cells. Some females will completely avoid clinical manifestations. Other females must pay some price for surviving gestation. For this reason, males and females have their own unique manifestations of identical mutations, and the female-only diseases are most often less severe than their male counterparts. Occasionally, however, females may manifest male-only diseases, either because loss of the mosaicism exposes deleterious genes or because the interactions have harmful effects. In addition, cellular mosaicism contributes to physiological diversity, because it provides an increased repertoire of expressed X-linked genes.